The Sequence Step Algorithm: A simulation-based scheduling algorithm for repetitive projects with probabilistic activity durations

The construction industry and academia have realized the critical path method and other time-based methods were not suitable for repetitive projects, which were resource-driven in nature. Both communities have been attempting to develop a better technique to schedule repetitive projects. Many approaches have been proposed; however, they are capable of solving the problems only to a certain degree of complication. Most of these approaches were limited to deterministic problems. A few probabilistic scheduling methods using simulation techniques were proposed with improvement in capturing the stochastic nature of construction activities; however, none of them guaranteed continuous resource utilization.;The Sequence Step Algorithm (SQS-AL) is a general scheduling algorithm for minimizing the duration of repetitive projects with probabilistic activity durations while achieving continuous resource utilization. SQS-AL consists of two main nested loops: the sequence step loop and the replication loop. For each sequence step, each replication loop is a simulation run that collects crew idle time for activities in that sequence step. The collected crew idle times are, then, used to determine resource arrival dates for user-specified confidence levels, i.e., probabilities of having zero idle time in corresponding activities. The process of collecting the crew idle times and determining crew arrival times for activities on a considered sequence step is repeated from the first to the last sequence step. The effect of scheduling activities on the crew idle times for following activities is revealed step by step prior to scheduling the following activities. As a result, SQS-AL can guarantee continuous resource utilization for the user-specified confidence levels.;This thesis also presents the application of work breaks, the determination of the controlling sequence, and the scheduling of resource-sharing activities in repetitive projects with probabilistic activity durations. An application, called "ChaStrobe," was developed on top of the Stroboscope Graphical User Interface to facilitate schedulers in creating simulation model for repetitive projects and scheduling the projects using all concepts presented in the thesis. In addition, ChaStrobe consists of two search methods, the exhaustive search and the genetic algorithm. Using the proposed concepts, the programmability in Stroboscope, and the search methods, ChaStrobe can optimize the scheduling problems of repetitive projects effectively.